Guided procedures, remote assistance and the field application around them — built phone-first, offline-first, and tested somewhere with mud on the floor. We build the app as well as the content, because a technician's job needs both.
The most common way an AR field service pilot dies is not technical. Practitioners report it repeatedly: the technician found it easier to make a video call than to put on a headset, so the headsets went in a drawer. That is a friction problem, not a rendering problem, and it should drive the hardware decision.
Our default recommendation for AR remote assistance and work instructions is the device the technician already carries and already charges. It is in their hand within seconds, it survives being put down, and nobody has to be trained on it.
There are genuine cases: both hands on the equipment, heavy gloves, a confined space, work at height. We build for headsets when one of those is actually true — and we will tell you when it isn't, before you buy the hardware.
Vaults, basements, plants and rural spans routinely have no usable signal. We cache the procedure, media and forms on the device so work continues with the radio off, then sync when coverage returns.
Where the work allows it, we deliver in the browser. No app store submission, no MDM ticket, no version drift across a thousand devices — you send a URL, and the update is live the moment we publish.
MadXR builds custom content of any kind — AR, VR, WebXR, web, mobile, websites and conventional software. For field work that usually means a guided procedure and the application it lives inside, delivered by the same team.
Step-by-step procedures with the relevant component identified on the live camera view or on a 3D model of the asset. One step at a time, confirmation where it matters, photo capture and sign-off at the points you need evidence.
Live video from the technician's phone with annotation, freeze-frame markup and pointer, so a remote expert can mark the actual valve rather than describe it. Session recording and notes attach back to the job.
Work orders, checklists, forms, photo and video capture, geolocation, sync, auth and an admin side. Most content studios cannot build this and most app shops cannot build the content. This is the seam we remove.
A chatbot or agent trained on your manuals, SOPs and service bulletins, so the answer comes from your revision rather than the internet — with a voice interface where the technician's hands are busy.
We would rather lose the AR line item than sell you a tool your crews abandon in month three. Several of these come up in nearly every scoping call.
If the task is six steps and the technician does it weekly, a one-page reference beats anything on a screen. It costs almost nothing, works in the rain, and never needs charging. Buy the card.
For "show me how this is done once", a well-shot three-minute video with a competent technician in it is faster to produce, cheaper to update and easier to watch. AR earns its cost when the procedure is long, high-consequence or done rarely.
Featureless housings, reflective steel, outdoor light that changes hourly — markerless tracking drifts in exactly those conditions. An overlay anchored to the wrong component is worse than no overlay. Sometimes the honest fix is a marker; sometimes it is a 3D model instead of a camera view.
Often the technician does not need an overlay. They need the current revision of the manual, yesterday's work order and the last three service notes, on one screen, offline. That is a cheaper build and a bigger result, and we will point at it if we see it.
Before anyone writes code you get an interactive storyboard — every step, every interaction, every line of voiceover, in a browser you can comment on. Your subject matter experts correct the procedure there, where a correction costs minutes.
A playable version with placeholder art, running on the phone or tablet your crews carry, in gloves, at the brightness of an actual work site. This is where most of the rework happens, deliberately.
We put it in front of the people who will use it, on site if you can arrange access. The questions we care about are whether they can reach the next step one-handed and whether it still works when the signal drops.
One structured round of user acceptance testing, then deployment, integration into your work management or LMS, and the source. Additional scope is quoted separately rather than absorbed silently.
We publish our prices. Almost nobody in this industry does, and it is the single most common complaint buyers have about it.
| Scope | Typical use | Price |
|---|---|---|
| Quick interactive experience | One procedure, a job aid, a proof of concept for a single task | $5,000 – $10,000 |
| Interactive platform | Multiple procedures with progress tracking and reporting | $10,000 – $15,000 |
| Chatbot trained on your documentation | Manuals, SOPs and service bulletins answered in the field | $6,000 – $12,000 |
| Custom AI agent | Agent that acts on your systems, not just answers questions | $15,000+ |
| Field application — iOS | Native app for a single platform, typically the fleet's iPads | $20,000 |
| Field application — iOS and Android | Cross-platform, for mixed or BYOD fleets | $30,000 |
| Tier 1 — simple work instruction module | Basic procedural task, up to 10 minutes | $30,000 |
| Tier 2 — task training module | Multi-step procedure, 15–20 minutes | $40,000 – $45,000 |
| Tier 3 — full scenario module | Comprehensive simulation, branching decisions | $50,000 – $60,000 |
Common add-ons: offline mode +$5,000 · video streaming +$7,000 · real-time chat +$6,000 · geolocation and maps +$3,500 · camera and photo capture +$3,000 · knowledge base and RAG document ingestion +$3,000 · voice interface +$4,000 · additional language +$2,000 · LMS, SCORM or SSO integration +$2,500 · analytics +$1,500 · tool and API integrations quoted against your systems. Terms are 30% at kickoff, 35% at UAT, 35% on final delivery, with one round of UAT included.
Utilities, energy, telecom, construction and manufacturing — the operators whose people work on equipment in places the office has never seen.
Fiber splicing documentation and field support for crews working on live infrastructure.
Water and sewer field operations, including hydrant work and CCTV pipe defect identification for inspection crews.
Utility truck operations training for line and service crews.
Journey management and hazard recognition, delivered on tablet and desktop for field and office use.
A browser-delivered site walk with an AI companion, built bilingually in English and Spanish for a construction environment.
Instrument servicing training for field service engineers, built as a proof of concept and then extended.
AR work instructions present a procedure step by step on a device the technician is holding or wearing, with the relevant part of the equipment identified on the live camera view or on a 3D model of the asset. Instead of reading step 14 in a PDF and then working out which of six identical valves it means, the technician sees the step and the part together. In practice most of the value comes from sequencing, confirmation and capture — one step at a time, a photo or a sign-off at the points that matter — rather than from the overlay itself.
Usually not, and we will say so before you buy any. Practitioners report the same failure again and again: an AR remote assistance pilot dies because the technician found it easier to make a video call than to put on a headset. Our default recommendation is the device the technician already carries — the phone or tablet in the truck. We recommend headsets only where hands-free genuinely matters: both hands on the equipment, heavy gloves, a confined space, or work where putting the device down is not an option.
Nothing custom deploys in a day, and any vendor who tells you otherwise is selling you a trial account. What genuinely deploys in a day is an off-the-shelf video-assist tool running on the phones your technicians already carry, or a browser-delivered tool you send as a link — no app store submission, no device management, no MDM ticket. The honest sequence is to prove the workflow with the simplest thing that works, find out where it stops fitting your dispatch, forms and asset data, and then build custom only for that gap. We are happy to be the second step and not the first.
MadXR publishes its pricing. A focused single-topic AR experience runs $5,000 to $10,000, and a multi-module interactive platform $10,000 to $15,000. A native iOS field application is $20,000, or $30,000 for iOS and Android. Full AR work instruction modules run $30,000 for a simple procedure, $40,000 to $45,000 for a multi-step task and $50,000 to $60,000 for a full scenario. A chatbot trained on your existing documentation is $6,000 to $12,000, and a custom AI agent starts at $15,000. Terms are 30% at kickoff, 35% at UAT and 35% on final delivery.
It has to, and that shapes the build from the start. Field sites routinely have no usable connectivity — underground, inside a vault, in a plant, or simply out of coverage — so we cache the procedure, media and forms on the device, let the technician complete and capture work fully offline, and sync when the device is back in coverage. Offline capability is an add-on at $5,000 because it is real engineering: conflict handling, queued uploads and storage limits, not a checkbox. Live remote assistance is the one thing offline mode cannot rescue, which is another reason to build the guided instructions the technician can use alone.
Yes, and this is the main practical difference between us and most vendors in this category. Content studios deliver a module that has to live inside somebody else's platform; software shops build the app but subcontract the 3D and the instructional design. A field technician's job normally needs both — the work order, the checklist, the photo capture, the sign-off and the sync, plus the guided procedure inside it. We build AR, VR, WebXR, mobile, web and conventional software in-house, so the same team delivers the whole thing and there is no seam to argue over.
Send us the procedure, the callback you keep paying for, or the job your newest crew gets wrong. We'll tell you whether AR is the right answer — including when a PDF would do.
Start a conversation